New SB Acoustics 3 way from Troels

Well, maybe the voicing works for his listening room but might not for others.

This is one of the drawbacks of building a kit designed by someone else, as it has been tuned in to his listening room.
For example, my room is not large but still needs 5-6dB compensation, say 100Hz starting point and I bet this SB 3-way will be lean in my room.

Finally, I continue to say that a published design sensitivity/baffle-step compensation should relate to an endless wall in terms of sensitivity rather than someone's room.
 
Last edited:
I consider 6dB (flat) perfect amount of BSC (200 in regard to 1000 hertz). That way, with good positioning of loudspeakers in room one can get good and balanced sound.

When you know it's flat anechoic then every problem you observe when measuring in room response is room induced and much of it is placement specific.
 
I consider 6dB (flat) perfect amount of BSC (200 in regard to 1000 hertz). That way, with good positioning of loudspeakers in room one can get good and balanced sound.

So how can you name enough baffle-step compensated for this speaker with 87dB? The infinite-baffle sensitivity of the woofer is 88dB, this is still only 1dB difference.

And why do builders report problems with tonal balance? Is this a coincidence?

Somebody wants to place the speakers far away from the walls or have a large room with little room gain.
 
Last edited:
Crossover frequency between woofer and midrange is at 200Hz. Wavelenght of that frequency is 1.7m so unless your loudspeaker is 1.7 meters or further of back wall, it will behave almost as on the infinite baffle.

Also, when crossing that low, it will usually form a hump bellow 100Hz so you can easily get excess energy lower. Notch on his woofer probably has some high value coil (>10mH), large capacitor and not large enough resistor which in turn makes the impedance in that place plummets to about three ohm - but he needed to do that to fix the frequency response bump. Looking at impedance phase, i'd say it is centered at circa 50Hz.

EDIT: I stopped reading his writings on his designs long time ago. I read it now and i see he wrote what he did with notch on woofer which is in line with what i wrote.

The SS sheet calculates effective response from a given baffle width. Excellent! From target baffle width we can count on some 85 dB sensitivity. From an "infinite" baffle we're close to 89 dB. Now, what the SS sheet doesn't do is adding the room-gain depending on speaker position and room dimensions. No wonder, these calculations depends on a lot of things, but from my other EXCEL sheet I get some 3-4 dB room gain from 55 cm above floor, 80 cm to front wall and 180 cm to side wall. Based on experience we can expect a system sensitivity i the area of 88-89 dB. Not bad from such small enclosure - and bass extension.
 
Last edited:
Crossover frequency is at 200Hz. Wavelenght of that frequency is 1.7m so unless your loudspeaker is 1.7 meters or further of back wall, it will behave almost as on the infinite baffle.

Ok, then maybe my room is very bad, but the closest wall to the speaker is 1.3 meters and the front wall is approx. 80 cm, but my speaker doesn't work like "almost" in an infinite-baffle.
 
Last edited:
How so, are you missing some bass or there is lots of it from 6dB bsc speakers ?

If you're having problems of that sort then welcome to the modal region of the room :) that can only be corrected by repositioning speakers, room treatment or multiple subwoofers or all three combined.

Andrew Jones - industry professional - worked for a whole day to position Elac Debut B6 on one of the hifi shows in US to get the tonal balance in his listening room. He is professional and needed a whole day until he was pleased. Why would it be any different for any listening room - but most of us aren't Andrew Jones so we can expect a bit more time moving the bastards around our rooms till we get the sound that is worth listening :)
 
Last edited:
25mm makes a difference somewhere, not elsewhere, but even if doing it is not sure that it is going in the right direction and that it is tied to a particular position which is not necessarily the best.

This is starting to go off-topic, but my room is just as if the lowest frequencies would be amplified only by room-gain, say below 60Hz.
By the way, I just wanted to point out that the speakers designed by TG do not necessarily work everywhere and I wrote an example of why and I don't think I would be alone.

My speakers thank you very much, they sound good because they work in the room where they were designed.
However, somebody else pays for the DIY plans but doesn't get help from the seller to fit into his room and just say that he needs to put on more carpets.
After all, I understand TG at some level, because there are so many variations, but as I wrote this is one of the drawbacks if you use someone else's DIY plans and you can't change it to your liking, in the absence of measurements.
 
I think you can. You can measure and put the resulting graphs into something like xsim and see what happens when you play, or you can play with eq on the computer and see that you want say 3db down at 1k and 5db down at 12 k, then again you can play. That is how I tune my speakers to the room. Ttoels designs are his best effort to fit any room, and will always be a poor compromise on what can be achieved.
 
Let's scrutinize that measurements to see where we're at.

View attachment 730411

If you look at 200Hz level and lower it's about 2dB hotter than 400Hz-1200Hz. In return, 400-1200Hz is also about 2dB hotter than 1200 and beyond.

What does this means ?

Since lower in frequency we go, we get less directivity untill the sound becomes omni, this means that the level of lower frequencies doesn't fall off axis as fast as it happens to higher frequencies. This makes 200Hz and lower to remain loud and becomes louder in regard to higher frequencies as you go off axis. That's where so much bass comes from. Midrange directivity is not constant also and it also falls off slower at lower frequencies in regard to higher. Same applies for the tweeter. All this leads to bass heavy, dark sound.

Now, you'd think that TG doesn't know what he's doing, right ? Wrong. It's not what he does but more where he does it. In Scandinavia wooden houses are quite usual, while in East Europe they are almost exclusively made out of brick and/or cinder block.

This kind of voicing is usual for lossy wooden walls that flex and absorb much of that energy down low or for a very large listening room. Put this speaker in a classic British or East Europen house/apartment made out of hard materials and add few walls in close proximity to the speakers and it will be mayhem with bass.

For frequency response that suites smaller rooms (under 30 sqm) made out of hard materials one should turn to Kef voicing. Not letting the woofer dig much lower than 35Hz is preferable. If you have really good listening room or just can afford to place the loudspeaker where ever you want in the room, then full 6dB baffle step design is the most you can afford to do, usually it gets to about 4dB. Troels did 8dB baffle step here and it shows on measurements and in listening also.

Looking at this measurement i could bet that Troels listened and tweaked this speakers in house with drywalls or wooden walls.

@Zvu
EXACTLY my thoughts!!! Thats why I'm listening with my bass port partially stuffed/sealed. Also when I was experimenting with DSP EQ I found out that even as little as 0.5 db attenuation is very notable i suppose due to room gain multiplier... So my plan is to try simple L pad on bass.
 
In my case, Troels was well responding until I have asked him what could be wrong with the tonal balance I had where he only suggested that I simply needed more damping material like carpets in my listening room. I already have carpet and furniture in my room. I have other speakers so I know something is wrong in the levels of mid/ low high frequency range. In my case, my issue is not fine tuning but more serious. I had my friend to come over and he has ATS-4 (also from
Troels ). He confirmed the too much amphasis on the mid/hi frequency range with music having higher energy in that range for my SBA-10. However, Troels indicated that voicing is done in the design and both ATS-4 and SBA-10 should have the same tonal balance response....

I have listened to Bassivus played audio files and they seem well reproduced. However, I would be curious that he compares with music having more energy in the mid/low hi frequency range at higher audio level...just to verify if I did something wrong in the build or the result is as designed. I have built the crossover using the Jantzen kit with detailed instructions, following reverse polarity for the mid driver, ...
In order to have a more enjoyable tonal balance I had to add a considerable resistor value in series with R1....but I know it is only a rough patch and can bring other imbalance...like reducing corner frequency of the tweeter even further down...

I am even wondering if some parts delivered by Jantzen are not in spec ...or there are variation in the manufacturing process of the tweeter so mine have greater sensitivity or higher gain in the low end...

I am even considering changing the mid/hi crossover with help from a friend since I have no experience in that area.

Any hint?

Fab

in your case i would consider another detailed check on the wiring of XO. If it turns out ok I would try putting damping material with masking tape on the baffle edges of upper part and maybe even pillow behind that "empty" baffle part and have a listen.
 
Padding down the woofer passively is not a good idea. Why not reduce the attenuation of the mid/tweet section?

tried it... sounded awful! At least on tweeter. I have reduced attenuation on mid slightly. You would'n believe how much room plays role in bass gain. when I was playng with placement in one place I had 20db peak on 65hz and just 10 cm from there almost flat
 
Yeah, another swamp of the passive xover, level matching is not easy often.

In the end I'm planning attenuating with L-Pad for less then 1db also improved the lowest point on impedance curve slightly. There are other problems there but I can not discuss it openly since I was more than once reminded that I must keep the XO values confidential :rolleyes:
However TG openhanded accepted my remark that R11 on bass is heavily underrated and said he will change it to 4 resistors in parallel in future kits... I burned my finger on it after playing slow sine sweeps on bass...
Of course that was before "I insulted" him
 

fab

Member
Joined 2004
Paid Member
in your case i would consider another detailed check on the wiring of XO. If it turns out ok I would try putting damping material with masking tape on the baffle edges of upper part and maybe even pillow behind that "empty" baffle part and have a listen.
I have checked my crossovers many times but maybe a different eye...
For the damping material you are the second person giving me that advise so I may proceed...
I have tried the pillows behind the tweeters but no change really.

Fab
 
@Zvu
EXACTLY my thoughts!!! Thats why I'm listening with my bass port partially stuffed/sealed. Also when I was experimenting with DSP EQ I found out that even as little as 0.5 db attenuation is very notable i suppose due to room gain multiplier... So my plan is to try simple L pad on bass.

L-pad on woofer, though feasible, is not particularly good idea. If i'd make some atenuation of that sort, i'd probably use series woofer coil for it. Rdc can be anything from 0.25-1 ohm. Any larger value than that and i'd do biamping.

I tried 4 ohm series coil for a specific subwoofer aplication some 10 years ago. It worked without a problem but the coil got warm (not hot) when playing really loud. You'd need heck of a resistors to make an L-pad that would be immune to that. Not to mention the decadency of spending so much power on fancy heater :)

Changing series rdc in front of woofer will affect the Q of the speaker but since you are using it sealed, it is negligible.

The smartest thing would be, in my opinion, to totally redesign the crossover or to use biamping: woofer on one amp, mid-hi on another with the existing crossover.
 
Last edited: